Android Safety & Ethics Researcher

We think a machine with a body is easier to hold to account than one without — you can watch it, it works at human speed, and it is subject to the same physical limits we are. We do not think a body makes anything safe by itself. Making that difference real is what this role is for.

It sits alongside engineering rather than after it, and it has standing to say no.

What you'll do

• Define the safety envelope for a machine that operates in physical contact with people, including how it behaves when it is uncertain.
• Design the evaluations that would surface a problem — adversarial testing and longer studies, rather than demonstrations that went well.
• Work on the parts of our evaluation standard concerned with how people actually experience and respond to this technology over time.
• Set the rules for how data about people is collected, retained and withdrawn, and make sure they are followed.
• Write down what we have decided not to build, and why.
• Publish methods and findings so people outside the company can check them, including when a result is unflattering.

What we look for

What we're looking for

• Research background in human-robot interaction, AI safety, applied ethics, human factors, or a related field.
• Empirical rigour — the ability to design a study whose result could genuinely go against what you expected.
• Willingness to hold a position under commercial pressure, and judgement about when the pressure is reasonable.
• Clear writing for a general audience, not only a peer-reviewed one.
• Interest in the specific questions a human-like machine raises: attachment, dependency, consent, and who is affected first.

Nice to have

• Experience with ethics review, or with safety standards for machinery that works alongside people.
• Work with people in care, clinical or educational settings.
• Policy or standards experience, particularly making a technical argument to a non-technical body.